Family support and medication adherence among adult type 2 diabetes: Any meeting point?

Background: Serious disability diabetes mellitus (DM) cause for patients and their support system-family and friends are enormous. It remains an important public health challenge, despite effective medical therapies for its management; patients' poor adherence remains a global problem. Objective: The study assessed the relationship between family support and medication adherence among adult Type 2 DM (T2DM) attending family medicine clinic of a rural tertiary hospital. Methodology: This was an analytic cross-sectional hospital-based study carried out among 367 patients selected by systematic random sampling method. Relevant data collected through a semi-structured questionnaire and clinical data recorded and fasting plasma glucose (FPG) were utilized for the study. Data were analyzed using SPSS version 16.
Results: The mean (standard deviation) age of respondents was 61.7 ± 11.4 years. Those with strong family support that achieved medium/high (Morisky Medication Adherence Scale-8 > 6) level of medication adherence (odds ratio [OR] [95% confidence interval (CI)] = 1 6.4 [9.1-29.6], P < 0.001) constituted 69.5% of respondents. Family support was also found to have a direct relationship to glycemic control (FPG < 7.1 mmol/l), 65.7% of those with strong family support achieved good glycemic control, P < 0.001, OR (95% CI) = 17.4 (9.2-37.2). The level of medication adherence was noted to be directly related to glycemic control, 79.4% of those with medium/high medication adherence had good glycemic control, OR (95% CI) = 25.0 (14.4-43.6), P < 0.001. Strong family support leads to higher medication adherence level which resulted into better glycemic control.
Conclusion: Family support improves medication adherence and glycemic control. Physicians should explore patients' family support system to improve medication adherence level and better management outcome of chronic diseases especially T2DM.

INTRODUCTION

Over the last two decades, the global rise in diabetes mellitus (DM) and the burden it place on sufferers as well as on the orthodox health-care systems they utilized for care have gathered critical attention in the international health community. In the sub-Saharan African context specifically, it was noted by researchers that not only has there been a sudden rise in prevalence, but morbidity and mortality rates of individual living with DM are on the increase with a huge financial and social costs implication to control and treat its complications.[123] Its physical and psychosocial consequences for patients and their family; and its economic impact on the society are enormous.

Diabetes management is quite complex and requires addressing many other issues beyond glycemic control. A situation where diabetes patients visit clinics regularly and their blood glucose levels still remain uncontrolled despite the treatment they receive is a challenge that calls for attention, to which a solution must be sought.[45] The targets of DM treatment are to maintain metabolic control, improve quality of life, and prevent complications. This basically comprises nonmedical and medical therapy, the nonmedical is related to behavioral changes associated with a healthy diet and physical exercise,[6] while medical therapy involves glucose control with hypoglycemic drugs. Findings from various studies demonstrate that poor glycemic control results in the disease progression, hospitalization, development of long-term complications, premature disability, and mortality.[7]

Medication adherence represents a fundamental concept in offering care effectively to DM patients, and the understanding of barriers and facilitators to behavioral changes can encourage interventions with a view to furthering or strengthening conditions that promote adherence. Pascal et al. recorded 72.5% adherence in SE Nigeria, while Adisa et al. recorded adherence rate of 60.2% in Ibadan SW Nigeria.[89] Many reasons has been alluded to influence the control of diabetes and of particular importance is psychosocial factors;[10] several of these pertain to family life and have been shown to correlate with metabolic control. Psychosocial factors are relevant to almost all aspects of its care, therefore, identifying the level and pattern of family support within a diabetic population will be an important step in improving quality of care.[711] Thus, family members involvement may improve the care of type 2 DM (T2DM). Family support is considered a complex and dynamic process that involves individuals and their family networks, working to satisfy their needs, provide and complement the resources they have and thus, cope with new situations.[2]

So much has been learnt from research work done in developed countries concerning psychosocial aspects of diabetes care, the same cannot be said of diabetes research in Nigeria and other developing countries. This study is designed to identify how perceived family support relates with medication adherence in our environment, with expectation that physicians can improve adherence to medication by their patients using its result.

METHODOLOGY

This was an analytic cross-sectional hospital-based study conducted from April to September 2016 in the Department of Family Medicine, Federal Teaching Hospital, Ido-Ekiti, Ekiti State, South-Western, Nigeria. This tertiary healthcare facility serves as a referral center for other health institutions (general hospitals and comprehensive health centers) in Ekiti State and environs. The hospital runs a postgraduate training in the department of family medicine and six others. The family medicine department is well staffed with its components of medical officers, registrars, consultants, and nursing staff and runs a daily diabetes clinic. The clinic is run by the resident doctors under the supervision of consultant family physician. The participants were diabetes patients aged 30 years and above, attended clinic at least twice in the past 6 months and on oral hypoglycemic agents (OHA) for more than 3 months, excluding those on combination of oral anti-diabetes medications and insulin, pregnant women and critically ill patients.

A sample size of 367 was statistically determined for the study using a medication adherence prevalence rate of 60.2% as reported by Adisa et al.,[9] the confidence interval (CI) of 95%, standard error of 5%, while putting into consideration 10% nonresponse rate, using the formula; n = z2p (1 − p)/d2 and

nf = n/[1 + n/N]. n = the minimum sample size when population is >10,000; nf = the minimum sample size when population is <10,000; N = the estimated population size in a year (T2DM in 2015 at the FETHI family medicine diabetic clinic = 3640).

Systematic random sampling technique was used to recruit participants among the diabetes patients attending the clinic. The family medicine clinic record of the hospital showed that between 14 and 17 patients with T2DM attended clinic daily which translated to about 75 patients per week using an average of 15 patients per day. Family medicine department runs outpatient clinic majorly 5 days in a week thus 300 patients were expected in a month. A total of 1800 diabetic patients were expected to be encountered during the 6 months of study. Hence, sample interval was calculated thus:

K = N/n where K = sample interval, N = total number of patients to be encountered, n = calculated sample size = 1800/367 = 4.9 (approximated to 5). This translated to recruiting every fifth diabetic patient who attended the clinic daily. Each day, the nurses were told to separate and collate folders of all T2DM patients who usually present early because of the need to do fasting plasma glucose (FPG) at each visit. The first five in order of presentation were then selected and made to pick from a container with five papers labeled with one “Yes” and four “No.” The patient who picked “Yes” became the starting point for the day. Selection of other respondents was based on the sample interval, which is 5. The patient's folder was marked to prevent recruitment of a patient twice. This process was repeated daily until 370 patients were recruited over 6 month (26 weeks).

Pretested semi-structured questionnaire drafted in English Language and translated in Yoruba (local language) and back-translated to English was used to obtain relevant information on the characteristics of respondents. The questionnaire was majorly administered in English and in Yoruba for a selected few who do not understand English.

Medication adherence was determined using Morisky Medication Adherence Scale-8 (MMAS-8). The MMAS-8 enquires about a patient's experiences with medications during the 2 weeks prior to answering the questionnaire.[12] This questionnaire has been translated into different languages so that it can be used with ease among patients in different settings. It is an eight item scale scored as follows: Items 1–4 and 6–8 are reverse-coded, i.e., a NO is scored 1 and YES is scored 0. Item 5 is direct coded a NO is scored 0 and a YES is scored 1 while item 8 has five options scored in a negative direction from 1 (never) to 5 (always). Item 8's score is further divided by 4 (1/4) when calculating a summated score. The total scale has a range of 0–8. The level of medication adherence was classified as low adherence (<6), medium adherence (6–7), and high adherence (8). For the purpose of this study, MMAS-8 scale was dichotomized into two categories, these being medium/high adherence (score ≥6 points) and low adherence (score <6 points). The MMAS-8 scale is reliable (Cronbach's α = 0.83), with a sensitivity of 93% and a specificity of 53%.[1314] The advantages of this over other methods of measurement include its simplicity, speed, low cost, and validity of use.

The level of family support was assessed with Perceived Social Support-Family Scale (PSS-Fa). The PSS-Fa is a t20-item validated measure of perceived family support developed by Procidano and Heller.[15] They reported an alpha coefficient of 0.9 indicating good internal consistency. PSS is the degree to which one perceives how his or her needs for support are fulfilled by family and friends. Subjects response “yes,” “no,” or “I don’t know,” with each “yes” answer scoring +1, “no” response scores 0 and “don’t know” −1. Items 3, 4, 16, 19, and 20 were reversed scored (a “no” response was scored as +1). Summated scores were used to arrive at a social support score. Possible range of scores is 0–20. Summated scores indicate levels of social support, which was rated accordingly as “strong” or “weak” social support. Scores below 10 was classified as weak family support, while scores 11 and above was classified as strong family support.

Five research assistants who were a resident doctor, a nurse, two information officers and a medical laboratory scientist were used. The research assistants were recruited and trained by the investigator for 2 days for the purpose of data collection, viz-a-viz informed consent process, questionnaire administration and accurate clinical data measurement.

FPG test was used to determine the level of glycemic control as an objective biomedical indicator of family support. It was assessed using mean value of FPG test results over the most recent three consecutive clinic visits. The first two were from patients’ record while the 3rd was done by researcher at no cost to the patient. Levels of control were classified as: poor control - ≥7.1 mmol/l, good control - ≤7.1 mmol/l.[16] Blood pressure (BP) was measured using an appropriate cuff-sized Accosson® mercury sphygmomanometer and stethoscope. The BP was recorded to the nearest 2 mmHg. Two measurements were taken after the patient has been allowed to sit for 5–10 min and an average value was taken and classified according to JNC 7.[17]

Surgifield Sm-160 Stadiometer made in England was used to measure the height of subjects in centimeters. The height was measured with subjects standing erect against the wall-mounted scale, wearing no shoes, and looking straight ahead. Measurements were taken and recorded by a trained research assistant nurse to the nearest 0.01 m. Weighing scale RGZ-160, by Lincoln Mark Medical, England, was used to measure in kilograms. A standard 10 kg weight was used daily to crosscheck accuracy of the weighing scale. The scale was regularly adjusted to correct for zero error at the beginning of each day and after each patient. Measurements were taken and recorded to the nearest single decimal place.

Data gathered were entered and analyzed using Statistical Package for Social Sciences 16.0 (SPSS Chicago Inc., IL, USA). Frequency tables were generated for relevant variables. Means, standard deviations (SDs), one way ANOVA, Tamhane's post ad hoc analysis, odd ratio and percentages were determined as appropriate. The means and SD were calculated for continuous variables while categorical variables were analyzed using proportions. Test of significance was done using Student's t-test. The Student t-test was used to compare the means when there were only 2 means to compare. P ≤ 0.05 was taken to be statistically significant.

Ethical approval was obtained from the institutions’ Ethical Review and Research Committee. Informed verbal and written consent was obtained from the participants before the administration of questionnaires.

RESULTS

Data from 367 properly completed questionnaires were analyzed. The mean age of respondents was 61.7 ± 11.4, and the proportion of male 182 (49.6%) to female 185 (50.4%) was nearly equal. Table 1 shows clinical characteristics of the respondents. The mean BMI of respondents was 25.45 ± 3.83 kg/m2, 168 (45.8%) had good glycemic control, 340 (92.6%) did not smoke, and 337 (91.8%) did not drink alcohol. Majority 339 (92.3%) had one or more co-morbidities, while nearly all 342 (93.2%) used two oral anti-diabetic medications. Table 2 shows that respondents with strong family support had medium/high adherence compared to those with weak family support. Hence, family support is significantly related to medication adherence (χ2 = 110.6, df = 1; P < 0.001), odds ratio (OR) (95% CI) = 16.4 (9.1–29.6). A higher proportion of respondents with strong family support 155 (65.7%) had good glycemic control compared to respondents with weak family support 13 (9.9%). The observed difference is statistically significant (χ2 = 105.5, df = 1; P < 0.001), the OR is very significant OR (95% CI) = 17.4 (9.2–37.2) showing very strong association between family support and glycemic control.

Table 1

Summary of clinical characteristics of respondents (n=367)

Variables	Frequency (%)
BMI
Underweight	13 (3.5)
Normal	140 (38.1)
Overweight	177 (48.2)
Obese	37 (10.1)
Mean±SD (kg/m2)	25.45±3.83
FPG level
Good glycemic control	168 (45.8)
Poor glycemic control	199 (54.2)
Mean±SD (mmol/l)	8.9±3.7
Blood pressure (mmHg)
Systolic (mean±SD)	134.3±16.6
Diastolic (mean±SD)	84.7±9.4
Smoking history
Yes	27 (7.4)
No	340 (92.6)
Number of cigarettes smoked per day (n=27)
≤5 sticks	7 (25.9)
>5 sticks	20 (74.1)
Alcohol intake
Yes	30 (8.2)
No	337 (91.8)
Alcohol intake/day (n=30)
1-2 bottles	11 (36.7)
>2 bottles	19 (63.3)
Number of comorbidities*
Two or more	119 (32.5)
One	220 (59.9)
None	28 (7.6)
Number of medication used£
Two	342 (93.2)
One	25 (6.8)

*Hypertension, nephropathy, retinopathy, peripheral neuritis, £Metformin, glimepiride, glibenclamide, pioglitazone. BMI=Body mass index, SD=Standard deviation, FPG=Fasting plasma glucose

Table 2

Relationship between level of family support, medication adherence, and glycemic control

Family support	Adherence		χ 2	P
	Medium/high (n=180)	Low (n=187)
Strong	164 (69.5)	72 (30.5)	110.6	<0.001
Weak	16 (12.2)	115 (87.8)
Family support	Glycaemic control		χ 2	P
	Good (n=168)	Poor (n=199)
Strong	155 (65.7)	81 (34.3)	105.5	<0.001
Weak	13 (9.9)	118 (90.1)

Table 3 shows a significant relationship between adherence level and glycemic control; χ2 = 161.324, df = 1; P < 0.001. The OR (95% CI) = 25.0 (14.4–43.6). Table 4 shows that respondents with strong family support have lower mean FPG 7.5 95% CI (7.2–7.9) compared to respondents with weak family support 11.3 (10.6–12.0). The difference is statistically significant t − 9.641, df = 202 and P < 0.001. Mean difference = −3.8, 95% CI (−4.58, −3.02). Table 5 shows a trend in which the higher the level of adherence the lower the mean FPG. The trend was statistically significant (F 64.429, P < 0.001). Post hoc analysis for between and intergroup variations were all significant (P < 0.001).

Table 3

Relationship between adherence levels and glycemic control

Adherence level	Glycemic control		χ 2	P
	Good	Poor
Medium/high	143 (79.4)	37 (20.6)	161.3	<0.001
Low	25 (13.4)	162 (86.6)

Table 4

Relationship between family support and fasting plasma glucose

Variables	n	FPG (mmol/l), mean (95% CI)	t	df	P
Family support
Strong	236	7.5 (7.2-7.9)	−9.641	202	<0.001
Weak	131	11.3 (10.6-12.0)

FPG=Fasting plasma glucose, CI=Confidence interval

Table 5a

Relationship between adherence levels and fasting plasma glucose

Variables	n	FPG (mmol/l), mean±SD	F	P
Adherence levels
Low	187	10.7±3.9	64.429	<0.001
Medium	126	7.3±2.7
High	54	6.2±0.8

SD=Standard deviation, FPG=Fasting plasma glucose

DISCUSSION

The demographic pattern of the respondents revealed that the mean age (SD) was 61.7 ± 11.4 years. Mean age of patients in this study was slightly higher than that reported by Chinenye and others (57.1 ± 12.3 years) in the multicenter diabetic care study in Nigeria.[18] In this study, no respondent was <40 years of age, though in the selection criteria, minimum age was 30 years. This is because T2DM is very rare before 30 years of age.[19]

The prevalence of strong family support was 236 (64.3%). This is higher than findings by Adetunji et al. who recorded 49% strong support,[20] and Wen et al. reported far lesser strong family support in a developed country.[21] This study assessed family support using PSS (family) scale, while Wen et al. assessed for diabetes specific family support using Diabetes Family Support Behaviour Checklist-II specific for diet and exercise. The difference in instrument may be a way to explain this difference. It may however be too simplistic to attribute the difference to the research instrument. As noted by Afolabi et al., in West Africa, the family plays a very important role in the society, as the structure is usually robust; a family is traditionally extended to include grandparents, grandchildren, uncles, aunties, nephews, and nieces.[22] The nuclear family may also be extended through the acquisition of more than one spouse (polygamy), or through the common residence of two or more married couples and their children, or of several generations connected in the male and female lines. This structure provides basis for good structural support in West African families. It may not be surprising then that such a high level of support was recorded in this study.

Of the 367 respondents 180 (49.0%) had high adherence level which is similar to 49.3% reported by Arulmozhi et al. in Puducherry, Southern India,[23] but higher than Ghanian study by Bruce et al. that reported 38.5% using the same MMAS in our study.[24] A study conducted in the Netherlands on refill adherence and polypharmacy among patients with type-2 diabetes in general practice showed that mean adherence with oral glucose lowering drugs was between 61 and 85%;[23] however, it may be difficult to compare with this work, because pill counting was used, while MMAS was used in this work.

In this study, medium/high medication adherence was associated with good glycemic control. This study positively compares with study done by Pascal et al. where blood glucose control was significantly higher among patients that adhered (72.5%) with their anti-diabetic medication compared with their low adherent counterparts.[8] A study by Yusuff et al. found 93% of respondents with good glycemic control were adherent to OHA.[25]

The benefits of good glycemic control have been documented in previous studies, these include prevention of acute illness and reduction of the risk of long-term complications.[26] Family support was significantly associated with glycemic control. The result is consistent with Adetunji et al. report that 49% of respondents had strong family support which was significantly associated (P = 0.008) with good (<7.0 mmol/l) FPG control.[20] This finding could be explained by the central role a family plays in patient management, especially in chronic illness like diabetes in our environment.

Comparison between the respondents with strong and weak support revealed higher medication adherence was achieved by those with strong family support. This may translate to enjoying a high degree of traditional social support from either members of his nuclear or extended family or both. In Africa generally and West Africa in particular, people are shielded from untoward effects of circumstances by support from families and friends.[27] A research done by Khosravizade et al. revealed that family support helps to improve medication adherence in patients with type 2 diabetes.[28] A strong family support is likely to be associated with high medication adherence because diabetes is regarded as a family disease and not that of individuals.[29] Adisa et al. recorded 13.2% family support, among other factors influencing medication adherence.[9] The finding of weak family support among 115 (87.8%) of low adherent respondents, partly supports similar studies where weak support behavior was predictor of not taking medication as recommended by a physician.[30] It is argued that diabetes must be understood not only by the individual who has developed it, but also by his/her family members and other significant people in their lives. Supports from loved ones and others regarding diabetes management program are vital throughout the life of people with diabetes.[212223] Family members can have a positive and/or negative impact on the health of people with diabetes, interfere with or facilitate self-care activities (e.g., reminding to take a prescription), and contribute to or buffer the deleterious effects of stress on glycemic control. Although family members can provide many kinds of social support (e.g., emotional, informational, and appraisal support), instrumental support has been most strongly associated with adherence to self-care behaviors across chronic diseases like DM.[31]

CONCLUSION

Family support improves medication adherence and glycemic control. It is recommended that family physician should explore patients’ family support system to improve adherence level and better management outcome of chronic diseases especially T2DM.

Recommendation

The patients’ social support should be explored by physicians who manage patients with type two DM. It should become a routine to ask patients about their medication use and where poor adherence detected patients should be counseled appropriately and their family support reviewed.

Limitations

The limitations of this study include its cross sectional design which made it difficult to determine causality. MMAS-8 used in determination of medication adherence is subject to recall bias and social acceptability bias.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

Table 5b

Tamhane’s T2 post hoc test for inter-group differences in fasting plasma glucose

Adherence level comparisons	Mean difference (95% CI)	SE	P
Low versus medium	3.407 (2.512-4.302)	0.373	<0.001
Low versus high	4.558 (3.819-5.297)	0.307	<0.001
Medium versus high	1.151 (0.519-1.783)	0.262	<0.001

CI=Confidence interval, SE=Standard error